Ionospheric TEC predictions over a local area GPS reference network

Single layer ionosphere models are frequently used for ionospheric modeling and estimation using GPS measurements from a network of GPS reference stations. However, the accuracies of single layer models are inherently constrained by the assumption that the ionospheric electrons are concentrated in a thin shell located at an altitude of about 350 km above Earth’s surface. This assumption is only an approximation to the physical truth because the electrons are distributed in the entire ionosphere region approximately from 50 to 1,000 km. To provide instantaneous ionospheric corrections for the real-time GPS positioning applications, the ionospheric corrections need to be predicted in advance to eliminate the latency caused by the correction computation. This paper will investigate ionospheric total electron content (TEC) predictions using a multiple-layer tomographic method for ionospheric modeling over a local area GPS reference network. The data analysis focuses on the accuracy evaluation of short-term (5 min in this study) TEC predictions. The results have indicated that the obtainable TEC prediction accuracy is at a level of about 2.8 TECU in the zenith direction and 95% of the total electron content can be recovered using the proposed tomography-based ionosphere model.